Output Explorer

Every prompt in the paper, and what each model wrote back.

Answer a question about a scanned page from its OCR text. Scored by exact match and ANLS against the accepted answers.

13 of 5,330 prompts

Nearby prompts. All 5,330 DocVQA prompts

PromptDocVQA · pagejxmk0226_3.png

What is point number “5” in the page?

OCR text of the page · 972 characters; the scanned image itself is not published
Format :
5:00
Introduction/why film/Assn.
unaffiliated moderator
5 :05
5:30
Film showing
Q & A, with one doctor authority
6:00
Refreshments
7:00
End
Press/Infor-
mation Kit:
A .
News release
B .
Collateral materials
C .
"Sugar in the Diet of Man"
D .
Film booking card
Cost/funding: To be discussed
4 .
Member Showings: Suggest coordination among members
to avoid duplication of effort in HQ cities.
A.
Employees
B .
Customer groups
C.
HQ area media groups
D.
E .
HQ area TV Placement
HQ area professional groups/nutrition, health,
education
F.
HQ area government and civic groups
G.
H .
HQ area high schools and colleges
Plant community VIP's/media, government, schools
Support materials: Basic prepared introduction,
Collateral materials, SA basic literature
5 .
PBS Film Library distribution/if accepted, 264
stations .
6 .
Showing by SA dietitians in cities where Sugar
Information program is functioning.
#
Source: https://www.industrydocuments.ucsf.edu/docs/jxmk0226
System prompt · identical for every setup
Answer the question using only the OCR text from a single document page. Return only the answer, with no explanation. Preserve the answer wording from the OCR text when possible.
Expected answer
PBS Film Library distribution/if accepted, 264 stations
Models
4 of 4 columns · click a model to add or remove it

Ours

Partial · ANLS 0.96

PBS Film Library distribution/if accepted, 264 stations .

57 characters14 tokens

Aux 2015

Wrong

5:00

4 characters5 tokens

PiT-FT 2015

Wrong

Empty response.

0 characters

ChronoGPT 2015

Wrong

5:00

Question:

What is the point number “5” in the page?

OCR text: Format : 5:00 Introduction/why film/Assn. Unaffiliated moderator 5 :05 5 :30 Film showing Q & A

168 characters64 tokens